Aims: This proposal is for a randomized trial of ASSIST (Assistance, Support and Self-health) initiated through skill training for family caregivers, combined with a qualitative analysis of caregiver appraisal and coping strategies before and after skill-training. Aim 1 is to test the effects of ASSIST on psychological, social and health-related quality of life (QOL) outcomes with family caregivers within and across two chronic illnesses of aging that are similar in age and rate of onset, loss of functional independence and downward trajectory: Alzheimer's disease and Parkinson's disease. AIM 2 is to describe caregiver appraisal and coping strategies over time in persons with and without ASSIST. AIM 3 is to explore the effect of ASSIST on the economic quality of life outcomes of caregivers. Significance: Given projected increases in the number of frail elders over the next two decades, it is crucial to determine the portability and costs of informal caregiver education and training programs across various chronic illnesses of aging. Analysis of caregiver appraisal and coping strategies will generate data for targeting these caregiver strategies with specific training interventions in the future. Methods: Caregivers from each chronic illness will be randomized into one of two groups. The ASSIST Group will receive a combination of in-home and telephone skill-training contacts. The Comparison Group will receive only social telephone contacts. Measurements will be obtained at baseline, after 3 months of weekly skill-training contacts followed by 3 months of bi-weekly, maintenance contacts (T2), and 12 months after baseline (T3). Data gathered through semi-structured interviews with caregivers will be used to describe caregiving appraisal and coping strategies. It is hypothesized that ASSIST-trained caregivers within and across the two chronic illnesses will have better quality of life outcomes and more varied appraisal and coping strategies than comparison group caregivers. Analysis: A repeated measures multivariate general linear model will be used to test AIM 1 hypotheses about changes in caregiving burden and caregiving preparedness (psychological QOL) and caregiver-care recipient mutuality and amount of caregiving assistance (social QOL) over time, while a univariate approach will be used for health-related QOL (Aim 1) and economic QOL outcomes (Aim 3). ASSlST and comparison group caregivers' appraisal and coping strategies will be described and compared (AIM 2). Out-of pocket caregiving expenditures will be compared for the groups as well as the costs of delivering ASSIST per unit of change in caregiving burden (AIM 3). A final sample size of 142 will yield a power of .81 for finding moderate multivariate effects, using p = .025.